Secondary Logo

Journal Logo

Cardiovascular Anesthesia: Research Report

The Marked Reduction in Mixed Venous Oxygen Saturation During Early Mobilization After Cardiac Surgery: The Effect of Posture or Exercise?

Kirkeby-Garstad, Idar MD; Wisløff, Ulrik PhD; Skogvoll, Eirik MD, PhD; Stølen, Tomas; Tjønna, Arnt-Erik; Stenseth, Roar MD, PhD; Sellevold, Olav FM MD, PhD

Author Information
doi: 10.1213/01.ANE.0000219589.03633.BF
  • Free

There is growing evidence that active rehabilitation, including early mobilization, improved pain control, and early enteral nutrition, can reduce postoperative morbidity and mortality (1). The “fast track” approach with early tracheal extubation has facilitated early postoperative mobilization in cardiac surgery. Cardiac function is transiently impaired after cardiac surgery (2). The relevance regarding the cardiac surgical patient’s tolerance to early postoperative mobilization is, however, uncertain. Most studies on postoperative cardiovascular function are old; using treatment modalities not commonly used today. In addition, they all evaluated cardiac function at rest rather than during exercise, which is today considered routine after coronary artery bypass grafting (CABG) (3–5).

In previous studies we found that mixed venous oxygen saturation (Svo2) was markedly reduced and that there was no compensatory increase in cardiac index (CI) during mobilization on the first and second days after cardiac surgery (6,7). Despite hemodynamic data indicating that myocardial function was postoperatively reduced, the Svo2 was unchanged and CI was increased after surgery (8). These apparent inconsistencies may depend on the complexity of the model. When a patient is mobilized out of bed, several changes occur simultaneously. The present study aims at elucidating the probable mechanisms involved. We investigated the effects of posture and low-level muscular work on CI and Svo2 separately, with special attention to the influence of cardiac surgery on these physiological responses. The specific aims were as follows:

1. To study the preoperative and postoperative response to graded supine bicycle exercise and passive standing in CABG patients.

2. To estimate the amount of muscle work involved in passive standing and hence its role in the reduction of Svo2 by comparing the increase in oxygen consumption (V̇o2) during passive standing with that associated with supine bicycle exercise.

Methods

The Regional Board of Ethics in Medical Research approved the study protocol. As exercise testing has not previously been performed this early after cardiac surgery, we included only male CABG patients with ejection fraction >50% (preoperative angiography) who were exercise tested to >100 W without ischemia, or who, if not tested, were in New York Heart Association class II. An anesthesiologist not involved in the measurements independently observed the electrocardiogram (ECG) and patient during exercise. Exclusion criteria were acute surgery, serious renal or lung dysfunction, arrhythmias, and New York Heart Association class IV. Patients were recruited after obtaining written informed consent.

The patients followed our standard regimen for CABG surgery. Medication, except aspirin, was continued up to surgery. All medication except angiotensin-converting enzyme (ACE) inhibitors/AII blockers was resumed the first morning after surgery. Patients were monitored with standard 5-lead ECG, radial artery catheter, and reflectance oximetry pulmonary artery catheter (rPAC) continuously monitoring Svo2. Anesthesia was induced with diazepam, thiopental, fentanyl, and pancuronium and maintained with isoflurane and fentanyl. For cardioplegia, we used St. Thomas crystalloid solution. Extracorporeal circulation used a standard membrane oxygenator, α-stat blood gas control, pump flow of 2.4 L · m−2, venous temperature 34°C–36°C, and mean arterial blood pressure (MAP) between 50 and 70 mm Hg. Propofol was used for postoperative sedation. Tracheal extubation was done when standard criteria were met. Chest drains were removed on day one, 2–3 h before the postoperative measurements. Details are previously described (6,7).

On the day of operation, before premedication, an Explorer ™ rPAC (Baxter Healthcare Corporation, Irvine, CA) and an arterial catheter were placed using local anesthesia. After 15 min of recovery the patients were subjected to “standing” and supine “cycling” (Angio ergometer; Lode, Groningen, The Netherlands). The order of these procedures was randomized by the institutional clinical trials office and there was a 15-min rest between them. During cycling, baseline resting values were obtained with the legs fixed to the bicycle pedals and physiological responses were recorded during 10 and 30 W workload. During standing baseline values were obtained during supine rest and the physiological response to passive standing was recorded. At each step we used 2 min of stabilization before the measurements were made. A full hemodynamic profile was obtained including CI recorded as the mean value of three stable readings from 10-mL saline injection at room temperature. V̇o2 was measured by indirect calorimetry using a MetaMax 3× apparatus (Cortex, Leipzig, Germany). Mixed venous and arterial blood samples were analyzed for hemoglobin (Hgb), blood gas tensions, oxygen saturation, pH, and lactate with the ABL 625 combined blood gas analyzer and multi-wavelength oximeter (Radiometer, Copenhagen, Denmark). The patients subjectively graded the intensity of work on the Borg scale (9). The whole procedure was repeated on the first postoperative morning in reverse order to control for any carry-over effect caused by standing or cycling.

The MetaMax 3× apparatus was used with a 30% oxygen in air mixture both preoperatively and postoperatively. The pressure transducers were zeroed and positioned as previously described (6,7). The ECG was continuously observed for ST deviations or arrhythmias. Criteria for acute myocardial infarction were a persistent new q-wave in the ECG with creatine kinase subunit MB >50 μg/L after 20 h or troponin-T > 0.460 μg/L after 48 h (10). Derived physiological variables were calculated using standard formulae.

Descriptive results are presented as mean with 95% confidence interval; mean ± standard deviation, or median with range, as appropriate. Each patient was measured under all conditions and thus served as his own control. Between-day differences were analyzed using repeated-measurements analysis of variance with paired Student’s t-tests as post hoc analyses. Within days measurements were subject to paired Student’s t-tests with Bonferroni correction. SPSS version 13 (SPSS Inc, Chicago, IL) was used.

To illustrate the dynamic aspects of the reaction to different workloads and posture, a linear mixed model (11) was fitted to Svo2, CI, heart rate (HR), stroke volume index (SVI), V̇o2, and oxygen delivery Ḋo2) as separate response variables. This model uses all observations and thus improves overall precision. It should be noted that the estimates may differ slightly from the descriptive results where cycling and standing are considered separately. The following fixed explanatory variables were considered through backwards selection: order of procedures (i.e., cycling first versus standing first), preoperative versus postoperative, standing versus resting, workload (0, 10 or 30 w). Parameter estimates (e.g., change in CI per watt of workload)—referred to in the text as model estimates—were obtained by restricted maximum likelihood (REML) using the nlme package of the R statistical software (12). P values <0.05 were considered significant.

We did not expect different preoperative and postoperative responses in CI and Svo2 to the change in posture. No data were available regarding the effect on Svo2 of supine exercise of 30 W. Based on previous observations of resting CI (8) and the fact that 30 W increases V̇o2 by 150% of the resting value (13), a preoperative increase in CI by at least 1 L · m−2 · min−1 was expected. A preoperative versus postoperative differential response in CI to exercise of approximately 30% was considered of clinical interest. In a previous study resting CI displayed a preoperative and postoperative sd of approximately 0.35 L · m−2 · min−1 (8). Given a similar sd, 15–17 observations are required (paired Student’s t-test, power = 0.9, two-sided α = 0.05) (14). We included 16 patients.

Results

Demographic and perioperative data are given in Table 1. Seven patients had hypertension, four had previous myocardial infarctions, and one had experienced a cerebrovascular event without sequelae. Fourteen patients received β-adrenergic blockers as anti-anginal medication, three were receiving angiotensin-converting enzyme inhibitors/AII blockers, and two received calcium channel blockers. The placement of invasive monitoring devices and preoperative baseline hemodynamic measurements were uneventful. No patient met the criteria for perioperative myocardial infarction. No patient received blood transfusions. All patients were in sinus rhythm and had adequate pain relief during preoperative and postoperative measurements. The postoperative studies were done in the morning 14.8–22.1 h after end of surgery (Table 1). One patient found cycling at 30 W for 4 min too strenuous after the operation and stopped after 10 W. All others completed all parts of the study. No signs of myocardial ischemia were observed during measurements in any of the patients.

Table 1
Table 1:
Perioperative Data

Resting SVI remained unchanged after the operation but HR and thus CI were increased (Table 2). CI increased with increasing workloads on both days, mainly as the result of an increase in HR of approximately 1 beat/min per watt of workload. Unlike HR, CI and SVI increased less during postoperative exercise (Fig. 1A, B and C).

Table 2
Table 2:
Hemodynamics and Oxygen Balance During Supine Exercise
Figure 1.
Figure 1.:
Hemodynamic response to exercise with workload in watt (w); curves and individual observations. Parameters are estimated from a linear mixed model. P < 0.05 for all explanatory variables, except for the second degree term regarding SVI (P = 0.07), which was retained to illustrate the general response. “Jitter” (random noise) is added to the points to enhance visual clarity. A, cardiac index (CI) as a function of workload. B, heart rate (HR) as a function of workload. C, stroke volume index (SVI) as a function of workload.

The central venous pressure (CVP) increased, whereas MAP and systemic vascular resistance index decreased from preoperative to postoperative rest. This may indicate increased preload and reduced afterload after surgery (Table 2). Bicycle exercise increased mean pulmonary artery pressure and pulmonary artery wedge pressure both preoperatively and postoperatively. The systemic vascular resistance index decreased with increasing workload both days. CVP did not change with exercise before but increased with increasing workload after the operation (Table 2). Before the operation, the right and left ventricular stroke work indexes (RVSWI and LVSWI) increased with increasing workloads, indicating an increase in bi-ventricular performance. After the operation no such increase was seen (Fig. 2).

Figure 2.
Figure 2.:
Left (B) and right (A) ventricular performance during supine bicycle exercise immediately before operation (preoperative) and on the first morning after coronary artery bypass surgery (postoperative). The ventricular stroke work indices are plotted against the respective filling pressures (mean value ± 95% confidence interval). Actual workload (watts) is given for each point (0 = rest). Note the different scales on the y-axis in panels A and B. A, right ventricular stroke work index (RVSWI) plotted against central venous pressure (CVP). B, left ventricular stroke work index (LVSWI) plotted against pulmonary capillary wedge pressure (PCWP).

The arterial oxygen content was reduced from 20 ± 1 mL/dL before to 14 ± 2 mL/dL after the operation (P < 0.05), mainly because of lower Hgb-values (Table 1) but resting Ḋo2 was unchanged as a result of the higher postoperative CI (Tables 2 and 3). The exercise-induced increase in Ḋo2 was, however, less after than before the operation. The relative increase in V̇o2 with exercise was similar preoperatively and postoperatively; a workload of 30 W increased V̇o2 by 113% ± 53% and 118% ± 61% respectively (NS). The more pronounced reduction in Svo2 during postoperative exercise thus mainly depends on the reduced Ḋo2 (Fig. 3A, B and C).

Table 3
Table 3:
Hemodynamics and Oxygen Balance when Standing
Figure 3.
Figure 3.:
Changes in oxygen balance in response to exercise with workload in watt (w); curves and individual observations. Parameters are estimated from a linear mixed model. P < 0.05 for all explanatory variables. “Jitter” (random noise) is added to the points to enhance visual clarity. A, mixed venous oxygen saturation (Svo2) as a function of workload. B, oxygen delivery (Do2) as a function of workload. C, oxygen consumption (V̇o2) as a function of workload.

Resting arterial lactate was slightly higher after than before operation. No change was seen during preoperative exercise. Postoperatively lactate increased from 1.1 ± 0.3 mmol/L at rest to 2.2 ± 0.8 mmol/L at 30 W exercise (P = 0.001) (Table 2). The Borg score indicated that exercise was considered less strenuous before than after the operation (Table 2).

The change from supine rest to passive standing increased HR (model estimate) by 7 bpm, with no difference between days. The modeled reduction in SVI with standing upright was 17 mL · m−2 before and 9 mL · m−2 after the operation (P = 0.000), with corresponding reductions in CI of 0.8 and 0.3 L · min−1 · m−2 (P = 0.001). Before the operation, the cardiac filling pressures (CVP and pulmonary artery wedge pressure) decreased and MAP increased with standing; after the operation pressures were stable (Table 3).

Preoperatively the reduction in CI when standing resulted in a significant reduction in Ḋo2. After the operation, Ḋo2 remained unchanged from rest to passive standing (Table 3). The model estimated increase in V̇o2 was 89 mL/min on both days, indicating no difference in V̇o2 during upright standing from before to after surgery. The absolute reductions in Svo2 were 16% before and 11% after surgery (P = 0.021), reflecting a similar increase in V̇o2 on both days while Ḋo2 was reduced preoperatively and stayed the same postoperatively.

Continuous hemodynamic readings from the patient monitor and the rPAC Svo2 printouts confirmed that our measurements were obtained during relatively stable physiological conditions. The V̇o2 values calculated from the invasive measurements were lower than those measured by indirect calorimetry but demonstrated a similar pattern of change within days (Tables 2 and 3). We used the indirect calorimetry measurements in our calculations to avoid mathematical coupling of data. The order of procedures (standing or cycling first) did not affect the measurements.

Discussion

The present study shows that the cardiovascular response to exercise is altered on the first morning after CABG. The hemodynamic data indicate that myocardial function may be reduced. Further, judged by the reductions in CI, Ḋo2, and Svo2 during upright passive standing, postural change has less impact on hemodynamics and oxygen transport after than before surgery.

The patients had anemia after surgery, which reduced the oxygen-carrying capacity of the blood and also reduced blood viscosity, contributing to changes in preload and afterload (15). As a result of the postoperative increase in HR, resting CI increased, maintaining Ḋo2. The increase in HR may be compensatory to anemia, may reflect altered loading conditions, or may result from vagal tone suppression after cardiac surgery (16). The trace amounts of anesthetics left on the first morning after surgery should not influence hemodynamics. The dose of morphine given for postoperative pain control was small (Table 1), probably causing only very limited physiological effects, and the patients had no pain.

Increased V̇o2 after surgery rendered Svo2 reduced compared with preoperative values. The increase in V̇o2 during supine graded exercise was similar before and after surgery. On both occasions, it was compensated for by a combination of increased Ḋo2 (through increased CI) and increased oxygen extraction (resulting in reduced Svo2). The increase in CI during exercise was smaller, however, and the decrease in Svo2 was more pronounced after the operation. Both indices suggest that the cardiovascular compensation to increasing workloads may be attenuated postoperatively (Fig. 1A, Fig. 3A).

Despite indications of increased preload and decreased afterload after the operation, SVI was unchanged at rest. This may indicate reduced cardiac function (Table 2). Roughly judged by the parallel increase in HR, the shifts in autonomic balance during exercise were similar before and after surgery (Table 2, Fig. 1B). Further, the similarity of changes in filling pressures and vascular resistance with increasing workload indicate similar alterations in cardiac preload and afterload during exercise on both days (Table 2). The smaller increase in SVI with postoperative exercise may therefore indicate reduced myocardial function (Fig. 1C), illustrated by the plots of right and left ventricular stroke work indexes against the respective filling pressures (Fig. 2).

It has been reported that resting arterial lactate is slightly increased after uncomplicated cardiac surgery (17). Whereas low to moderate intensity exercise did not increase lactate before surgery, it significantly increased lactate postoperatively. This phenomenon has not been described previously. However, the study was not designed to investigate lactate metabolism; thus explanations will at this point be speculative.

Other studies have led to the opinion that the myocardial dysfunction after cardiac surgery is transient, in most cases reversed on the first morning after surgery (3–5). The exercise testing revealed indications that the postoperative myocardial dysfunction may be of longer duration than previously assumed. Echocardiographic studies have indicated that right ventricular dysfunction may last for months after cardiac surgery (18,19). The clinical significance of these findings regarding global cardiac function is, however, uncertain (20).

Standing up reduced the filling pressures before the operation. This suggests that reduced venous return is an important factor for the reductions in CI and Svo2 when the patients were standing. Invasive data were published confirming reduced CI and Svo2 during passive head-up tilt in normal subjects (21). However, we observed no changes in filling pressures with passive standing after the operation. This may depend on alterations in intravascular volumes, blood viscosity (15), or cardiovascular reactivity.

Passive standing demands muscular work. One may estimate the amount of this work by dividing the increase in V̇o2 when standing by the calculated increases in V̇o2 per watt during bicycle exercise. Dividing the modeled increment in V̇o2 of 89 mL/min when standing by the respective model increments in V̇o2 of 24 mL/min and 27 mL/min per watt when cycling (Fig. 3C) yields a workload of about 4 W. Entering 4 W into the equations in Figure 3A leads to estimated absolute reductions in Svo2 of 4% before and 5% after surgery. The muscular work during passive standing can be assumed to reduce Svo2 accordingly. The observed absolute reductions in Svo2 when standing of 16% and 11% led us to ascribe approximately 12% before and 6% after the operation to the change in posture per se. The estimated preoperative reduction of 12% is in good correlation with the study of Harms et al. (21), showing a reduction in Svo2 of 12% during sustained 70 degrees head-up tilt in normal subjects. The observed reductions in Svo2 of 16% preoperatively and 11% postoperatively are less than the corresponding reductions of approximately 18% and 19% obtained during preoperative and postoperative mobilization (8). This was probably a result of the lesser workload of passive standing.

Reduced Svo2 values at rest during the first hours after cardiac surgery have been correlated with a negative outcome (22,23). The Svo2 is inversely correlated to oxygen extraction and its level during exercise might represent the adequacy of the cardiovascular response to the muscular activity. Thus the determination of a “critical value” of Svo2 (i.e., the level at which anaerobic metabolism occurs) during postoperative exercise could be of clinical interest. However, there are very few data on the relations between oxygen extraction reserve and Svo2. From data given in a paper on the Everest II study (24), we calculated that a mean oxygen extraction rate of >82% (i.e., Svo2 <18%) occurred at maximal exercise in young, fit men. In a study on sedentary hypertensive men, the mean Svo2 at maximal exercise was 35% with an individual range from 20%–50% (25). Our aim for testing was to map the physiological changes corresponding to mobilization of postoperative cardiac surgery patients. We therefore did not test our patients at maximal exercise. The Borg score indicated that 30 W corresponded to 60%–70% of our patients’ maximal capacity (9). The finding of a mean Svo2 of 46% during exercise might support this. Combining V̇o2-values during mobilization of CABG patients in our previous studies (7,8) and V̇o2 data from graded supine exercise in this study leads to the assumption that postoperative mobilization represents a workload of approximately 10 W. In the present study, the patients were tested to 30 W without ischemia or negative outcome. Our preliminary conclusion is therefore that postoperative mobilization represents a low level exercise well tolerated early after cardiac surgery.

There are several limitations to the present study. The number of patients included is too small to draw definite conclusions regarding clinical implications of our findings. The workload during the supine bicycle exercise was very moderate. The study group had a relatively good preoperative exercise capacity and was slightly younger than our CABG patient population (for the year 2004, median age was 66.8 years; range, 38-85 years). Compared with previous results from CABG patients the level of Svo2 was higher whereas Hgb and CI were similar in the present study (7). Our results are therefore not directly applicable to all groups of cardiac surgical patients or higher workloads. Figures 1 and 3 illustrate the main aspects of response to bicycle exercise but extrapolation beyond the measurements is not justified. A more precise description would require measurements at more than two levels of workload. Further, more precise estimates of the effect of posture per se might have been obtained through a full factorial model that would require measurements during upright bicycle exercise. However, another exercise period on the first postoperative morning was not considered appropriate.

In conclusion, this study demonstrates that the activity-induced increase in CI and Ḋo2 is attenuated the first morning after CABG. Within the range studied, this altered response to exercise resulted in a more pronounced decrease in Svo2 with increasing workloads. Further, Svo2 was significantly reduced during passive standing versus supine rest through reduced CI. This effect was less pronounced after surgery, possibly because of altered loading conditions. However, the change in posture per se accounted for a substantial part of the reduction in Svo2 during preoperative and postoperative passive standing. We therefore conclude that the physiological responses to exercise and posture are both modified early after cardiac surgery and may both significantly contribute to the marked reduction in Svo2 during early postoperative mobilization.

The authors are grateful to the ICU nursing staff for their enthusiastic help.

References

1. Kehlet H, Wilmore DW. Multimodal strategies to improve surgical outcome. Am J Surg 2002;183:630–41.
2. St Andre AC, DelRossi A. Hemodynamic management of patients in the first 24 hours after cardiac surgery. Crit Care Med 2005;33:2082–93.
3. Roberts AJ, Spies SM, Sanders JH, et al. Serial assessment of left ventricular performance following coronary artery bypass grafting: early postoperative results with myocardial protection afforded by multidose hypothermic potassium crystalloid cardioplegia. J Thorac Cardiovasc Surg 1981;81:69–84.
4. Mangano DT. Biventricular function after myocardial revascularization in humans: deterioration and recovery patterns during the first 24 hours. Anesthesiology 1985;62:571–7.
5. Breisblatt WM, Stein KL, Wolfe CJ, et al. Acute myocardial dysfunction and recovery: a common occurrence after coronary bypass surgery. J Am Coll Cardiol 1990;15:1261–9.
6. Kirkeby-Garstad I, Sellevold OF, Stenseth R, et al. Marked mixed venous desaturation during early mobilization after aortic valve surgery. Anesth Analg 2004;98:311–7.
7. Kirkeby-Garstad I, Sellevold OFM, Stenseth R, Skogvoll E. Marked mixed venous oxygen desaturation during early mobilization after coronary artery bypass surgery: no predictive effect of preoperative ejection fraction Acta Anaesthesiol Scand 2005;49:827–34.
8. Kirkeby-Garstad I, Stenseth R, Sellevold OFM. The postoperative myocardial dysfunction does not affect the physiological response to early mobilization after coronary artery bypass grafting Acta Anaesthesiol Scand 2005;49:1241–7.
9. Borg GA. Borg’s perceived exertion and pain scales. Champaign, Illinois: Human Kinetics, 1998.
10. Lehrke S, Steen H, Sievers HH, et al. Cardiac troponin T for prediction of short- and long-term morbidity and mortality after elective open heart surgery. Clin Chem 2004;50:1560–7.
11. Everitt B, Rabe-Hesketh S. Linear mixed models I. In: Analyzing medical data using S-PLUS. New York: Springer, 2001:243–68.
12. R Development Core Team 2004. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, Available at: http://www.R-project.org. Accessed March 26, 2006.
13. Åstrand PO, Rodahl K. Textbook of work physiology. New York: McGraw-Hill, 2003.
14. Machin D, Campbell MJ, Fayers PM, Pinol APY. Sample size tables for clinical studies. Oxford: Blackwell Science, 1997.
15. Licker M, Ellenberger C, Sierra J, et al. Cardiovascular response to acute normovolemic hemodilution in patients with coronary artery diseases: assessment with transesophageal echocardiography. Crit Care Med 2005;33:591–7.
16. Bauernschmitt R, Malberg H, Wessel N, et al. Impairment of cardiovascular autonomic control in patients early after cardiac surgery. Eur J Cardiothorac Surg 2004;25:320–6.
17. Maillet JM, Le Besnerais P, Cantoni M, et al. Frequency, risk factors, and outcome of hyperlactatemia after cardiac surgery. Chest 2003;123:1361–6.
18. Wranne B, Pinto FJ, Hammarstrom E, et al. Abnormal right heart filling after cardiac surgery: time course and mechanisms. Br Heart J 1991;66:435–42.
19. Alam M, Hedman A, Nordlander R, Samad B. Right ventricular function before and after an uncomplicated coronary artery bypass graft as assessed by pulsed wave Doppler tissue imaging of the tricuspid annulus. Am Heart J 2003;146:520–6.
20. Hedman A, Alam M, Zuber E, et al. Decreased right ventricular function after coronary artery bypass grafting and its relation to exercise capacity: a tricuspid annular motion-based study. J Am Soc Echocardiogr 2004;17:126–31.
21. Harms MP, van Lieshout JJ, Jenstrup M, et al. Postural effects on cardiac output and mixed venous oxygen saturation in humans. Exp Physiol 2003;88:611–6.
22. Svedjeholm R, Hakanson E, Szabo Z. Routine SvO2 measurement after CABG surgery with a surgically introduced pulmonary artery catheter. Eur J Cardiothorac Surg 1999;16:450–7.
23. Polonen P, Hippelainen M, Takala R, et al. Relationship between intra- and postoperative oxygen transport and prolonged intensive care after cardiac surgery: a prospective study. Acta Anaesthesiol Scand 1997;41:810–7.
24. Sutton JR, Reeves JT, Wagner PD, et al. Operation Everest II: oxygen transport during exercise at extreme simulated altitude. J Appl Physiol 1988;64:1309–21.
25. Reybrouck T, Fagard R. Gender differences in the oxygen transport system during maximal exercise in hypertensive subjects. Chest 1999;115:788–92.
© 2006 International Anethesia Research Society